Amazon has escaped an ad ban from the Advertising Standards Authority (ASA), after the UK watchdog investigated claims that a TV ad for the AI assistant had unwittingly placed an order for cat food on behalf of a viewer at home.
The investigation marks the first time that the organisation has ruled on an accidental activation of a voice assistant device by a TV ad.
The ASA looked into a single complaint that the ad was "socially irresponsible" after receiving a report alleging that a film promoting the Echo Dot had activated a viewer’s own product and placed an order with Alexa for Purina cat food.
The commercial was broadcast in October 2017, however the complaint was not upheld by the ASA, which said that it was satisfied that Amazon “had taken measures to prevent its ads interacting with decides that might ‘overhear’ them,” and that the ad was not socially irresponsible.
In its response to the ASA, Amazon said it engineered the Echo Dot to prevent the device reacting to its own ads. The company noted that if this preventative technology failed, customers were to confirm purchases in order for them to be activated; meaning that an effective order could not be placed by an ad alone.
In the complainant’s case, the customer was able to immediately cancel the inadvertent Purina order.
AI assistants - of which Echo is the market leader - have featured in a series of bizarre news stories since they became widely available. Last year, Burger King proved that it was possible to ‘hack’ Google Home devices by running an ad capable of activating the smart home product. The tactic was lampooned during an episode of animated comedy South Park, in which the show succeeded in triggering viewers’ devices with a series of gag product orders.
Earlier in 2017, Amazon Echo customers in the US found their devices had embarked on unauthorised shopping sprees after tuning into TV ads, and in March, audio files recorded by an Echo device were submitted as evidence in an Arkansas murder trial.
Geraint Lloyd-Taylor, deputy head of advertising and marketing at law firm Lewis Silkin, told The Drum that it has become commonplace for AI assistants to be programmed to ignore TV ads.
“The real problem is that it’s much harder for manufacturers of this kind of device to guard against ads created by third parties; if it had been an ad for a cat food brand which deliberately said ‘Alexa; add XYZ cat food to my basket’, there’s not much Amazon can do to proactively guard against that.”
“From now on I would expect Clearcast to be on the lookout for obvious triggers. If it becomes a real problem (especially if advertisers decide to use this as a deliberate tactic) we might find that the ASA or even the Competition and Markets Authority (CMA) decide to issue guidelines or take more decisive action,” he added.
“If deliberate, consumers aren’t likely to stand for this kind of aggressive commercial practice by brands either, so I don’t think there’s a practical risk of it being adopted as a marketing practice.”
An Amazon Spokesperson reiterated to The Drum that to shop with Alexa, customers have to order the product, then confirm the purchase with a 'yes' response via voice. Any accidental purchases can be prevented by simply saying 'no'.
"Customers can also manage their shopping settings in the Alexa app, such as turning off voice purchasing or requiring a confirmation code before every order," Amazon added: “Additionally, orders placed with Alexa for physical products are eligible for free returns.”